Sony a7R II vs III
Moderators: rjlittlefield, ChrisR, Chris S., Pau
-
- Posts: 137
- Joined: Sun Jun 10, 2018 3:54 pm
- Location: Russia
Agreed - but that's not the concern. You're thinking of the final output, I'm referencing leeway to post-process for display or print too. If you need to pull detail out of shadows in a scene with bright highlights, the dynamic range of the source can easily cause a noticeable difference in the final image - even on an 8-bit monitor. 14-bits gives you 4x more leeway than 12-bits to "stretch" the range of shadow tones, or highlights, before visible banding or posterisation appears. Controlling the light is key, of course, but there are some pretty "high dynamic range" insects out thereJustwalking wrote:Small comment about examples with 12 and 14 bitsBeatsy wrote: I've tested that and can't find any significant differences between 12-bit and 14-bit output - or compressed vs uncompressed RAW (an option on all the high-end Alpha cameras).
All .jpeg on our monitors are only 8-bits. )
Here's a recent example. It's a Woolly Aphid at 5x, 6mm FoV. The Boris Johnson of garden pests
I gave up on it in the end, mostly because it was near impossible to pose well so important features were visible (like head and eyes for one). But the dynamic range between the white fluff and dark body is huge, to the extent that I didn't have leeway to pull sufficient detail into both (IMO). I probably should have used EFCS to get 14-bit source files - but given the other non-photogenic qualities, I didn't bother.
Lou Jost wrote:
Are there any cameras that disable their flash for EFCS?
Pau, are you sure about that? Here is a test by Jim Kasson comparing Nikon synch speed with and without EFCS:Canon and Nikon do
https://blog.kasson.com/the-last-word/f ... -revisted/
Flash is disabled in silent shutter mode in Canon and Nikon.
Maybe a 10bit monitor can help you, visually, to pull the details out of shadows or highlights during post. But then again, if final media or image format do not support it (>8bits), it is indeed useless, as those details will be lost.Beatsy wrote: Agreed - but that's not the concern. You're thinking of the final output, I'm referencing leeway to post-process for display or print too. If you need to pull detail out of shadows in a scene with bright highlights, the dynamic range of the source can easily cause a noticeable difference in the final image - even on an 8-bit monitor. 14-bits gives you 4x more leeway than 12-bits to "stretch" the range of shadow tones, or highlights, before visible banding or posterisation appears. Controlling the light is key, of course, but . . .
- rjlittlefield
- Site Admin
- Posts: 23564
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
See also http://www.photomacrography.net/forum/v ... hp?t=37163 regarding flash and EFCS on Nikon D850.
--Rik
--Rik
Exactly - hence the need to have leeway to stretch contrast in the source images so it can be mapped onto the (more limited) dynamic range of the typical output device. I am actually getting a 10-bit monitor soon, but that's to help with print proofing - not solely for final display per-se.mjkzz wrote:Maybe a 10bit monitor can help you, visually, to pull the details out of shadows or highlights during post. But then again, if final media or image format do not support it (>8bits), it is indeed useless, as those details will be lost.Beatsy wrote: Agreed - but that's not the concern. You're thinking of the final output, I'm referencing leeway to post-process for display or print too. If you need to pull detail out of shadows in a scene with bright highlights, the dynamic range of the source can easily cause a noticeable difference in the final image - even on an 8-bit monitor. 14-bits gives you 4x more leeway than 12-bits to "stretch" the range of shadow tones, or highlights, before visible banding or posterisation appears. Controlling the light is key, of course, but . . .
And just a quick note to point out I didn't mention these "extremes" in my original post about comparing 12 to 14-bit DR. That assumed light was under control. I brought it up now to illustrate the point that wide DR is still useful even if the typical monitor (or printer, or eyes) can't support it.
-
- Posts: 137
- Joined: Sun Jun 10, 2018 3:54 pm
- Location: Russia
Beatsy, i'm so sorry, but on my Dell P-series it's still 8-bit We can see only jpeg here.Beatsy wrote: Here's a recent example. It's a Woolly Aphid at 5x, 6mm FoV. The Boris Johnson of garden pests
From the sRGB standard, the white point is 80 cd/m^2 while the black point is 0.2 cd/m^2, which gives us a ratio of 400:1 or 8.6 stops.
-
- Posts: 137
- Joined: Sun Jun 10, 2018 3:54 pm
- Location: Russia
Pau wrote:Canon and Nikon doLou Jost wrote:Are there any cameras that disable their flash for EFCS?
Well, what I quickly wrote this morning with the phone seems OK, at least for Canon EOS (where Silent mode with LV means EFSC, there isn't full electronic shutter for now)Lou Jost wrote: Pau, are you sure about that? Here is a test by Jim Kasson comparing Nikon synch speed with and without EFCS:
https://blog.kasson.com/the-last-word/f ... -revisted/
Flash is disabled in silent shutter mode in Canon and Nikon.
With Canon xD and xxD models if the camera detects a EX flash it fires it and maintains LV and MLU but it disables EFSC, if you use a non EX compatible flash it just doesn't fire.
With Nikon the situation seems more complex, so please ignore Nikon on my post
Pau
Which is ideal for displaying an 8-bit (per colour channel) jpg. As I tried to explain in my last post - I am NOT talking about dynamic range of the final image that you view on your monitor. I am talking about dynamic range of the captured image(s). Higher dynamic range gives you more freedom to stretch areas of low contrast into a much wider range (relative to all the other tones in the image) before banding or posterisation becomes a problem.Justwalking wrote:Beatsy, i'm so sorry, but on my Dell P-series it's still 8-bit We can see only jpeg here.Beatsy wrote: Here's a recent example. It's a Woolly Aphid at 5x, 6mm FoV. The Boris Johnson of garden pests
From the sRGB standard, the white point is 80 cd/m^2 while the black point is 0.2 cd/m^2, which gives us a ratio of 400:1 or 8.6 stops.
-
- Posts: 137
- Joined: Sun Jun 10, 2018 3:54 pm
- Location: Russia
Point with noise is very interesting. You're right with EA ofcourse. Thanks.Lou Jost wrote: ..The result would not be the same, though, because the Oly takes eight shots rather than four, reducing noise more....
I think the pixel-shifting is easier, and on the latest Oly cameras it may also be faster.
Whatever EA you need for the pixel-shift, wouldn't you also need that same EA in twice-higher m telecentric lens in your approach?
Ahh, gotcha. But if there had been a problem in processing due to dynamic range, you would have seen it in the jpg. It's just that there wasn't any problemJustwalking wrote:I understand You, Beatsy. It was just my regrets that i can't compare differents trough .jpgBeatsy wrote: As I tried to explain in my last post - I am NOT talking about dynamic range of the final image that you view on your monitor. I am talking about dynamic range of the captured image(s).
-
- Posts: 137
- Joined: Sun Jun 10, 2018 3:54 pm
- Location: Russia
- rjlittlefield
- Site Admin
- Posts: 23564
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
JohnyM wrote:It's way eazier to get telecentric lens and shoot panorama on higher magnification IMO, also overall faster. Am i wrong?
Lou Jost wrote:Whatever EA you need for the pixel-shift, wouldn't you also need that same EA in twice-higher m telecentric lens in your approach?
If I understand Lou's point, then I disagree with it.Johnnym wrote:You're right with EA ofcourse.
The required aperture on the subject side would be the same in both cases, because you're shooting for the same resolution on subject.
But back at the sensor, the corresponding EA would only be half as wide with the pano setup. One way to see this is to apply the standard formula that Feff = m/(2*NA), doubling m while leaving NA unchanged.
It's also worth noting that in the pano setup the objective only has to cover half the subject dimensions at any one time, so it is simple to choose a higher magnification objective and pick up additional resolution on subject. That would come at the cost of deeper stacks, of course.
--Rik